What's your relative date filter set for? If you're using a static "minutes" value (say, 60), then depending on when your alert runs, you could miss alerts that run in the gaps. You can confirm that's what's happening by looking at the log files from VizAlerts and determining when it ran. If it was scheduled for 10am but ran at 10:05am, then it would potentially be missing extracts that failed between 9am and 9:05am (assuming the last execution ran on time).

If that's the problem, then to mitigate, I suggest you change to doing "previous hour", rather than specifying the number of minutes. That way, so long as your alert runs successfully, at any point over the following hour, it'll properly pick up any failures that occurred during the last hour (with no duplicate alerts either!).

If, however, the logs are showing that emails did go out to the proper people at the proper times, but the recipients say they didn't receive them, then it's time to go through rules, junk folder, "clutter", etc...and potentially follow up with whoever manages your SMTP server.

No problem! As a reward for being a loyal VizAlerts user , here's a bonus calc that will accomplish the same "safe relative date" logic, but for alerts on a 15-minute schedule:

Date calc which determines the proper date range of the last 15-minute interval as defined by the quarter hour mark, rather than simply a set number of minutes earlier than “now”:

// Get the start date for the previous quarter hour (the end date will be the same thing but without the first and last lines)

    DATEADD('minute', -15,

        DATEADD('minute',

            (INT(DATEPART('minute', NOW()) / 15)) * 15,

            DATETRUNC('hour', now()) )

    )

Edit: I don't know what I was on before. Replaced the "4"s with "15"s. I'll update again if I find any more issues...